摘要
根据已有成果设计了地方政府债务风险预警指标体系,在此基础上,吸收灰色关联法和BP神经网络在数据分析上的优势,建立了基于灰色关联理论-BP神经网络算法集成的我国地方政府债务风险指数预警模型。同时选用我国西部、中部、东部地区在2007-2009年期间的27个样本数据,运用该方法对已有地方政府债务风险进行分析,研究结果显示,地方政府债务风险呈现出不断上升的趋势。经检验基于灰色关联—BP神经网络指数预警方法具有计算时间短、客观性强、预测精度高,有较好的推广应用价值。
According to the previous achievements, risk warning index system of local govern- mental debt has been designed. Based on the system, grey relevance method and advantages of BP nerve net in data processing has been absorbed to establish the warning model of risk index of local governmental debt. 27 pieces of sample data from 9 counties in the eastern, middle and western parts of China from 2007 to 2009 have been applied to analyze the risk of local governmental debt through the above method. The result proves that China~ s local governmental debt risk in recen years. Tests show that the index warning method based on grey relevance and BP nerve net has high value of promotion and application, with short calculating time, strong objectivity and high predicating accuracy.
出处
《现代财经（天津财经大学学报）》
CSSCI
北大核心
2012年第7期96-104,共9页
Modern Finance and Economics：Journal of Tianjin University of Finance and Economics
基金
国家自然科学基金青年基金资助项目(71103060)
教育部人文社会科学研究一般项目(09YJC790077)
